Parameter Estimation of Frequency Hopping Signals Based on Numeral Charateristics in Compressed Domain and Atomic Norm

Abstract

The methods for parameter estimation of the frequency hopping signal(FHS) in compressed domain find the numeral chara-teristics of the compressed sampling data based on the measurement matrix, having a drawback of high computational complexity and a problem of base mismatch. We propose a method for parameter estimation of the FHS based on numeral charateristics in compressed domain and atomic norm. We Firstly establish a block diagonalization measurement matrix to piecewise compress signals, analyse the numeral charateristics of compressed sampling data to roughly estimate hop timing, and then separate signal segments without frequency hopping for estimating hop frequency precisely using atomic norm minimization. On the basis of the implementation of hop frequency estimation, design the dictionary of atoms and the hop timing precise estimation is realized in compressed domain. Simulation results show that the proposed methods frequency estimation performance is better than that based on grid CS, when the SNR is higher than −2dB and the compression ratio is higher than 0.5. The proposed method can also estimate precisely hop timing of FHSs.